Recent advances in miniaturization have led to the development of microfluidic systems that are designed, in part, to perform a multitude of chemical and physical processes on a micro-scale. Typical applications include analytical and medical instrumentation, industrial process control equipment, and liquid and gas phase chromatography. In this context, there is a need for devices that have fast response times to provide very precise control over small flows as well as small volumes of fluid (liquid or gas) in microscale channels. In order to provide these advantages, it is necessary that the flow control devices be integrated into the microfluidic systems themselves. The term “microfluidic” refers to a system or device having channels or chambers that are generally fabricated on the micron or submicron scale, i.e., having at least one cross-sectional dimension in the range from about 0.1 μm to about 500 μm. Examples of methods of fabricating such microfluidic systems can be found in U.S. Pat. No. 5,194,133 to Clark et al., U.S. Pat. No. 5,132,012 to Miura et al., U.S. Pat. No. 4,908,112 to Pace, U.S. Pat. No. 5,571,410 to Swedberg et al., and U.S. Pat. No. 5,824,204 to Jerman.
Although there are numerous micro-fabricated valve designs that use a wide variety of actuation mechanisms (Shoji and Esashi, J. Micromech. Microeng., 4, 157-171, 1994), most dissipate relatively large amounts of power to the chip or substrate or require complex assembly which limits their use in practical systems. Most microvalves are manufactured from silicon and are therefore not easily integrated into non-silicon microchip platforms such as silica, glass, or synthetic materials such as organic polymers. A microvalve using an electromagnetic drive is described in U.S. Pat. No. 5,924,674 issued to Hahn et al. Jul. 20, 1999. Microvalves using thermopneumatic expansion as the actuation mechanism and a shape memory alloy diaphragm and bias spring are commercially available. However, these microvalves suffer from the fact that they consume relatively large amounts of power during operation, typically between 200 and 1500 mW depending upon the design. This high power consumption can be a significant disadvantage when heating of the fluid must be avoided, when batteries must supply power, or when the microvalve is placed on a microchip. Moreover, valves using the aforementioned actuation mechanisms can only generate modest actuation pressures and consequently, hold off only modest pressures. Perhaps most importantly, these valve designs can be difficult and costly to manufacture and assemble, frequently requiring assembly in a clean room environment.
Recognizing that the power requirements of conventional valves limited their use in practical systems, Beebe et al. (Nature, 404, 588-590, April 2000) describe a flow control system consisting of a hydrogel. The hydrogel valves provide local flow control by expanding or contracting when exposed to various pH levels. While eliminating the need for associated power supplies, these valves suffer from slow response times (≈8-10 sec) and are able to withstand only modest pressure differentials.
Unger et al. (Science, 288, 113-116, April 2000) describe an arrangement for controlling fluid flow in microchannels. Flow control is accomplished by the use of soft elastomer “control lines” that intersect the microfluidic channels fabricated in an elastomeric substrate material. Applying pressure to the external surfaces of the control lines causes them to deform closing off that part of the channel they intersect. While eliminating the problem of power dissipation to the substrate, these valves require a microchannel having a specially shaped cross-section to seal properly. They also intrinsically require that pressure greater than in the channel be applied to the control line to keep the valve shut.
Ramsey in U.S. Pat. No. 5,858,195 provides for valveless microchip flow control by simultaneously applying a controlled electrical potential to an arrangement of intersecting reservoirs. The volume of material transported from one reservoir to another through an intersection is selectively controlled by the electric field in each intersecting channel. In addition to the need for elaborate switching and control of electrical potential, there are problems with leakage of fluid from one channel to another through the common intersection because there is no mechanical barrier to diffusion. Further, this flow control method has essentially no control over pressure-driven flow. For example, the flow control of a 10 mM aqueous buffer at pH 7, using a 1000 V/cm electric field in round channels about 50 μm in diameter, can be completely disrupted by a pressure gradient of only 0.1 psi/cm. Higher electric fields are generally prohibited because of rapid ohmic heating of the fluid. Furthermore, the presence of pH or conductivity gradients within the fluid can disrupt this valving scheme (Schultz-Lockyear et al., Electrophoresis, 20, 529-538, 1999).